Abnormality Pattern Detection Using Multi-Outlier Similarity Analysis

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing abnormality detection methods based on outlier detection techniques can only perform binary judgments of normal or abnormality and struggle to identify specific abnormality patterns, leading to difficulties in determining the type of abnormality and requiring engineer examination, as different methods detect different patterns and lack comprehensive coverage.

Innovation Solution

Calculating the similarity between deviation sensitivity information and deviation inclination information for each abnormality pattern to suitably detect and specify the type of abnormality, using a system that integrates multiple outlier detection methods and stores sensitivity information systematically.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If multiple outlier detection methods are used to improve abnormality detection coverage, then detection capability is improved, but device complexity increases

Engineering Contradiction:
Improveabnormality detection capabilityVSAvoidsystem complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent segments the abnormality detection task by dividing it into multiple specialized outlier detection methods, each targeting specific abnormality patterns. Instead of using one complex method to detect all abnormalities, the system divides the detection space into multiple segments (different abnormality patterns) and assigns specialized detection methods to each segment, thereby improving overall detection capability while managing complexity through structured division

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent creates a universal abnormality detection system that integrates multiple outlier detection methods into a single platform capable of detecting various types of abnormalities. The system performs multiple functions (different detection algorithms) within one unified framework, allowing it to handle diverse abnormality patterns without requiring separate systems for each detection method

Inventive Principle:
Principle #6Universality (Multi-functionality)

2Ease of operation

If binary judgment method is used to simplify abnormality detection, then ease of operation is improved, but loss of information increases

Engineering Contradiction:
Improvedetection simplicityVSAvoidabnormality pattern information
Core Design Contradiction:
Ease of operationVSLoss of information

Solution Approach 1:

The patent segments the abnormality information by categorizing it into distinct abnormality patterns. Instead of providing a single binary judgment, the system divides the information output into multiple pattern categories, each representing a specific type of abnormality. This segmentation preserves information about the nature and type of abnormalities while maintaining operational simplicity through structured classification

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent transitions from a one-dimensional binary judgment (normal/abnormal) to a multi-dimensional classification system that adds the dimension of abnormality pattern types. By introducing this additional dimension, the system preserves rich information about different abnormality patterns while maintaining ease of operation through automated pattern recognition and classification

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Data Source

PatentEP3795975B1Abnormality sensing apparatus, abnormality sensing method, and abnormality sensing program
Publication Date: 2023.08.02 MITSUBISHI ELECTRIC CORP
  • EP3795975B1 patent drawingFigure 1
  • EP3795975B1 patent drawingFigure 2
  • EP3795975B1 patent drawingFigure 3

AI summary

A deviation inclination calculation unit (22) calculates a deviation score by using evaluation data obtained from a subject apparatus as an input, in each of a plurality of outlier detection methods specifying data deviated from other data from among subject data, and calculating deviation scores indicating deviation degrees of the data specified, and calculates deviation inclination information from the deviation scores calculated. An abnormality detection unit (23) calculates, for each abnormality pattern, a similarity degree between deviation sensitivity information indicating sensitivity for each of a plurality of abnormality patterns with respect to each of the plurality of outlier detection methods, and the deviation inclination information calculated, and detects an abnormality of the subject apparatus.